The progressivity of personal income tax in South
Africa since 1994 and directions for tax reform
T.J. Steenekamp
3
ABSTRACT
The imperative to reduce poverty levels in South Africa requires not
5
less but more public expenditure and probably higher tax levels. The
purpose of this study is, firstly, to examine the impact of personal
income tax reforms since 1994 on the tax structure and its scope to
meet the challenges of rising needs and equity. Secondly, the study
aims to provide a broad outline of personal income tax approaches
that are globally considered as alternative reform options. The study
finds that in South Africa, direct taxes as a percentage of total tax
revenue increased in importance between 1993/94 and 2010/11.
The personal income tax burden for wage earners in South Africa
has remained fairly constant since 1995. The personal income
tax structure is progressive, but there was a declining trend in
progressivity between 1994 and 2009. Increasing personal income
tax rates is constrained by low company tax rates, possible increased
efficiency costs and ‘herd behaviour’. The income tax system in
South Africa conforms to a semi-comprehensive income tax system.
The tax reform option that holds the most promise for developing
countries (and South Africa) is the dual income tax system.
Key words:
6
personal income tax, tax reform, progressivity, tax burden, marginal tax
rates, tax threshold, comprehensive personal income tax, dual income tax
Introduction
Tax systems evolve over time in a quest to improve economic performance and to
finance social goals. Various factors – some of which are internal to countries and
others of which are global – impact on tax reforms in individual countries. The South
1
Prof. T.J. Steenekamp is in the Department of Economics, University of South Africa. E-mail: [email protected]
Southern African Business Review Volume 16 Number 1 2012
39 T.J. Steenekamp
African economy is characterised by high unemployment, large social and income
inequalities and low levels of foreign direct investment. Economic growth rates are
at levels that are unlikely to reduce poverty significantly over the short term and even
longer term. To achieve all the UN Millennium Development Goals (MDGs), major
investments are required that need to be financed (UN Millennium Project 2005).
The impact of the global financial crisis of 2007/08 led to huge bailouts and rescue
plans for governments, which resulted in budget deficits and bloated public debt to
GDP ratios. South Africa was less affected, but not spared, and faces fiscal austerity
measures that will hamstring expenditure and/or lead to tax increases.
The UN assumes major increases in government expenditures on the MDGs of up
to four percentage points of GDP until 2015 and suggests using a broad-based revenue
source such as a value-added tax. The Organisation for Economic Cooperation and
Development (OECD) (2010a: 20–21) concluded that the best tax policy options for
promoting economic growth (from a least-distortionary perspective) are recurrent
taxes on immovable property, then taxes on consumption followed by other property,
personal income tax and lastly corporate income taxes. Globalisation tends to put
pressure on developing countries to lower their levels of taxation (Tanzi 1996,
2004). We note, however, that the imperative to reduce poverty levels in developing
countries, including South Africa, requires not less but more public expenditure and
probably higher tax levels. These opposing pressures force policy-makers to reduce
inefficiencies in public expenditure and at the same time to change the tax system to
generate sufficient revenue in an equitable manner. Tanzi (2004: 534) remarks that
the two ‘work horses’ that must carry this burden are the value-added tax (VAT) and
personal income tax. VAT was introduced in South Africa in 1991 and has become
what is generally considered to be an efficient revenue ‘milch cow’ but regressive
tax. Personal income tax serves the purpose of raising revenue and ensuring that
equity objectives are reached. Since high-income earners generally benefit most from
globalisation, the personal income tax system is ideally suited to capture revenue
from these income groups for redistributive purposes (Tanzi 2004: 540).
In this study, the focus will, firstly, be on the impact of personal income tax reforms
in South Africa since 1994 on the tax structure and the scope of personal income tax
to meet the challenges of rising expenditure needs and equity. Secondly, the study
aims to provide a broad outline of personal income tax approaches that are globally
considered as alternative reform options. Trends in the composition of taxes before
1994 and changes in the direct tax (income tax) mix in South Africa since 1994 are
reviewed. The personal income tax (PIT) burden, the progressivity of the PIT rate
structure and the distributional impact of PIT reforms are analysed. The potential of
PIT rate changes and reform alternatives in the design of the tax system to address
40 Personal income tax reform in South Africa since 1994
the revenue and distributional challenges are briefly reviewed. Of particular interest
are the findings of two comprehensive reports on tax reform in the United Kingdom
(IFS 2010, 2011; OECD 2006, 2010a). A developing country perspective is obtained
by reviewing the research of eminent fiscal economists on tax policy and reform in
developing countries.
Tax composition and trends in direct (income) taxes
since 1994
Direct (income) tax trends in the post-1994 period cannot be viewed without
considering trends in the composition of taxes in the preceding years. Figure 1
provides some perspective and shows how the composition of the main sources of
tax revenue (as a percentage of total tax revenue) changed between 1981 and 1994
(financial year ending March). It is observed that corporate income tax (CIT) and
personal income tax (PIT) exhibit a scissor-like pattern, with PIT increasing from
a low 16.6% in 1980/81 to 39.6% in 1993/94. In contrast, CIT decreased from 39.7%
to 13% over the same period. At the same time, VAT as a percentage of total revenue
increased from 12% in 1980/81 to 26.3% in 1993/94. PIT and VAT are ultimately
borne by persons, and if one were to add the incidence of the fuel levy and other tax
types such as user charges, it becomes clear that the relative tax share of individuals
in total tax revenue increased significantly from the early 1980s to the early 1990s.
Against the backdrop of a rising personal income tax burden, the Katz Commission
was tasked with a comprehensive review of the tax system in South Africa and released
its first interim report in 1994 (Katz Commission 1994). This report was followed
by another nine reports between 1994 and 1999. In giving direction to tax reform,
the Katz Commission at the outset recognised that the high fiscal burden had to be
addressed by paying attention to reducing most marginal individual rates, reducing
the direct personal income tax burden and nominal rates of corporate tax, increasing
the tax base, and improving tax collection (Katz Commission 1994: 11). The Katz
Commission (1994: 12) was also astutely aware that political and equity considerations
precluded a shift to higher VAT rates and instead bargained on improving economic
growth rates and other policy interventions to increase tax revenue.
Figure 2 shows revenue trends in personal income tax (PIT), company income tax
(CIT) and direct taxes (PIT and CIT) since 1993/94. Direct taxes as a percentage of
total tax revenue increased by 10.5% between 1993/94 and 2010/11. Using three-year
moving averages, the increase was 9.8%. A comparison reveals that OECD countries
(18 of the 30 countries) have also increased their share of direct taxes since 1996
(OECD 2010a: 28). In contrast, developing countries have increased their share of
1
41 T.J. Steenekamp
1
1
1
Source: SARB (2011b)
Figure 1: Main sources of tax revenue as a percentage of total tax revenue (1980/81
to 1993/94)
70.0
60.0
50.0
PIT as % of total tax revenue
Percentage
40.0
CIT as % of total tax revenue
30.0
Real GDP growth
20.0
Direct tax revenue as % total tax
revenue
10.0
2010/11
2009/10
2008/09
2007/08
2006/07
2005/06
2004/05
2003/04
2002/03
2001/02
2000/01
1999/00
1998/99
1997/98
1996/97
1995/96
1994/95
-10.0
1993/94
0.0
Source: National Treasury 2011 (Annexure B: Table 2); SARB 2011a (S148)
Figure 2: Direct taxes (PIT and CIT) as a percentage of total tax revenue (1993/94 to
2010/11)
42 Personal income tax reform in South Africa since 1994
indirect taxes from 34.9% in 1990 (25.5% in 1970) to 40.1% in 2000 (see Bahl & Bird
2008: 282). It seems that the tax mix in South Africa is more aligned to patterns in
developed countries.
A further inspection of the trends in the relative importance of PIT revenue and
CIT revenue to total tax revenue is revealing. Between 1993/94 and 1998/99, the
contribution of PIT far exceeded the CIT share, and these ratios remained fairly
stable until 1998/99. This period coincided with the economic downswing that lasted
from December 1996 until August 1999. From 1999/2000, a converging pattern
emerged. PIT revenue as a percentage of total tax revenue declined as real GDP
growth increased until at least 2006/07. However, CIT revenue as a percentage of
total tax revenue increased over the same period and seems to be in tandem with the
upturn in the economic cycle that lasted from September 1999 to November 2007. As
the economy grew, CIT revenue increased dramatically (although from a low base)
by over 450% in nominal terms compared to the approximate 64% increase in PIT
revenue. When the economy started slowing down in 2008, income tax revenues
continued to increase, but from 2009/10 the share of CIT decreased sharply as CIT
tax revenue started to drop in nominal terms. Overall the share of PIT in total tax
revenue decreased somewhat from 39.6% in 1993/94 to 34.1% in 2010/11. In contrast,
the share of CIT almost doubled from 13% to 22.7% over the same period. Income tax
shares in the OECD area exhibit a similar pattern between 1995 and 2007 (OECD
2010a: 28).
Changes in the income tax shares are largely endogenous, but some may be
attributed to policy interventions such as the successes of the South African Revenue
Service (SARS) in broadening the tax base and sharpened administrative vigilance,
which impacted on compliance and tax rate reductions. In the case of company tax
revenue increases, there may be a Laffer-type response to lower company tax rates.
Since 1994, the personal income tax system was changed in many respects (see
National Treasury and Department of Finance [various issues]; Manuel 2002) and
in response to the recommendations of the Katz Commission. The number of tax
brackets was reduced from ten to six; the child rebate was scrapped; the individual
is now the unit of taxation; and the primary rebate was increased annually to
compensate for inflation and to retain progressivity. South Africa also moved from
a source of income principle to the worldwide base of taxation. The tax base was
broadened by making almost all fringe benefits taxable and by taxing capital gains
since 2001. Capital gains tax from individuals netted government R4.4 billion in
revenue in 2009/10 (approximately 2.1% of taxes on persons and individuals).
1
43 T.J. Steenekamp
The personal income tax burden and progressivity in
South Africa
What was the impact of the post-1994 reforms (including rate reductions) on the
personal income tax burden, the progressivity of the PIT schedule and the distribution
of personal income?
1
The personal income tax burden
A measure of the average (effective) personal income tax rate (or burden) is obtained
using national accounting data (see SARB 2011a). In 1995, current taxes on income
and wealth of households in South Africa were R51 623 million. If the gross balance
of primary income of households as defined in the national accounts is used as the
tax base (R398 605 million), the personal income tax burden was 13.0%. In 2010, the
burden came to 13.5%. Over the 15-year period, it reached a maximum of 14.2% in
1999 and recorded a low of 11.3% in 2004. This narrow definition indicates that the
personal income tax burden has remained fairly constant since 1995.
The burden of personal income tax can also be expressed as a percentage of GDP.
Calculated from IMF (2008), the unweighted average central government PIT/GDP
ratio for South Africa was 8.4 in 2007. This compares to a PIT/GDP ratio of 8.2 for a
selection of 13 developed countries and 2.0 for a selection of ten developing countries.
However, once social security contributions are added (10.1) to the personal income
tax revenue in developed countries, the PIT (including social security contribution)/
GDP ratio increases to 18.3.
Drawing conclusions on the basis of these tentative comparisons on the potential
for additional PIT revenue must be done with caution. South Africa is a developing
country and has a limited tax base of approximately six million registered taxpayers
(National Treasury and SARS 2011: 6). These taxpayers must support close to
14 million recipients of social assistance grants. If government proceeds with its
intended course of making social security contributions by employees mandatory,
it is quite clear that the personal income tax burden of South African taxpayers will
rise significantly and will be even more out of line with developing country levels
and patterns. In an attempt to estimate and compare tax performance, Steenekamp
(2007) used the representative tax system approach to compile a tax effort index. The
results for the financial years 2000/01 to 2004/05 indicate that the South African
revenue authorities outperform comparable countries. When the different tax classes
were considered, it was concluded that South Africa uses personal income tax and
corporate income tax very intensively and that the total tax burden is high. These
results suggested lower personal income taxation and corporate taxation (Steenekamp
1
44 Personal income tax reform in South Africa since 1994
2007: 14). As noted, the relatively small income tax base and the sharp increases in
user charges on traditional public services such as education, toll fees and security on
the overall tax burden of individuals must be acknowledged. It then becomes quite
apparent that the scope for additional revenue from this source is indeed limited.
The progressivity of PIT
Declining top rates of personal income tax in the past two decades and more is
one of the most common global occurrences in taxation. As the OECD (2010a: 33)
points out, however, this has not been matched by a reduction in the average income
tax that is levied on the income of the average production worker. How has this
impacted on progressivity in the personal income tax structure? Various methods
can be used to measure progressivity. Nyamongo and Schoeman (2007) considered
a number of such methods and reported empirical results for South Africa based
on the procedures developed by Musgrave and Thin (1948) and Kakwani (1977).
Regarding the reform period after 1994, Nyamongo and Schoeman’s study produced
mixed results. According to the Musgrave and Thin procedure, progressivity declined
between 1994 and 2003/04, whereas the Kakwani disproportionality measure shows
that progressivity increased ‘impressively’ between 1994 and 1999/2000, but then
declined in response to tax reforms between 1999/2000 and 2003/04 (Nyamongo &
Schoeman 2007: 487–489).
Progressivity is affected by marginal rates but also by what has happened to
thresholds. In this study, three measures are employed to determine how adjustments
to the PIT rate and thresholds affected progressivity. All three measures compare the
marginal and average tax rates for an individual worker at the average wage level in
the economy (excluding agriculture and the public sector). This is shown in Table 1
for South Africa and a selection of OECD countries, along with the OECD average
(unweighted). For the industry categories noted in Table 1, the average annual
nominal wage for South Africa was calculated at R49 014 in 1997, R66 499 in 2000
and R123 676 in 2009.
The first progressivity measure considers the threshold at which the top marginal
rate applies. In 1997, taxpayers on average had to earn just twice the average wage
before the top rate became applicable (column 1). This multiple increased to 3.6 times
in 2000 and to almost five times in 2009 – an indication of less progressivity. This is
in contrast to the OECD pattern. The income threshold (as a multiple of the average
wage), where the top statutory PIT rate for the OECD is levied, declined between
2000 and 2009.
1
45 T.J. Steenekamp
Table 1: Income tax rates of average workera (2000 and 2009, and 1997 for South
Africa)
Top threshold
(multiple
of average
wage)b
Marginal
tax rate
Average tax
rate (100%
of average
wage)c
Average tax
rate (67%
of average
wage)
Average tax
rate (167%
of average
wage)
Progressivity
(1)
(2)
(3)
(4)
(5)
(4) ÷ (5)
2000
2009
2000
2009
2000
2009
2000
2009
2000
2009
Australia
1.2
2.8
44.5
31.5
26.1
22.0
20.7
15.6
34.5
28.3
Hungary
0.9
0.8
56.0
45.0
23.2
21.2
17.5
11.9
30.3
Korea
5.5
3.2
6.6
14.5
2.1
4.2
0.7
1.3
6.4
49.3
4.7
12.8
10.8
0.8
4.0
-5.9
-1.7
Mexico
2000
2009
60.0
55.1
27.7
57.8
43.0
8.7
10.9
14.9
7.9
11.3
-74.7
-15.0
New Zealand
1.7
1.5
21.0
34.0
19.4
18.4
18.6
15.6
24.2
24.9
76.9
62.7
Poland
3.3
2.8
9.3
8.8
6.6
6.4
5.3
5.2
7.7
7.4
68.8
70.3
Turkey
OECD
average
South Africa
(1997)
8.1
3.0
17.8
17.6
14.7
12.2
13.2
9.5
18.0
15.5
73.3
61.3
4.4
(2.9)d
2.6
28.9
27.2
16.4
15.4
12.3
10.7
23.1
22.0
53.2
48.6
3.6
4.8
37.0
18.0
18.0
10.1
11.2
6.2
27.0
15.8
41.4
39.4
(2.0)
(41.0)
(18.0)
(9.2)
(27.8)
(33.2)
Notes:
a An adult full-time worker in the covered industry sectors whose wage earnings are equal to the average wage
earnings of such workers.
b Top PIT threshold as multiple of the average wage.
c Gross average earnings are calculated using a broad set of industry sectors including categories C to K inclusive as defined in the International Standard Industrial Classification of All Economic Activities (ISIC Revision
3.1, United Nations).
d Excluding Mexico.
Sources: StatsSA (1998, 2003, 2011); National Treasury, Budget Review (various issues); OECD (2010b)
Table 1 shows that both the marginal tax rate (column 2) and average tax rate
(column 3) for the average earner in South Africa declined between 1997 and 2009.
This trend is also observed in some of the OECD countries, but was not that drastic
considering the 1.0% decline in the average tax rate in all OECD countries. A second
measure of progressivity that can also be used here is to calculate the difference
between the marginal tax rate (column 2) and average personal income tax rates
(column 3), divided by 1 minus the average personal income tax rate, for an average
worker (see OECD 2010a: 34 & 44). A higher number indicates higher progressivity at
the earnings level of an average worker. In 1997, the percentage was 0.28, compared
to 0.08 in 2009 implying that progressivity declined over this period. In 2009, the
OECD average was 0.14, with Turkey at 0.07 and Hungary at 0.30. However, this
progressivity measure does not consider the impact of tax changes on lower and
higher incomes.
46 Personal income tax reform in South Africa since 1994
The third progressivity measure compares the burden faced by persons earning
two-thirds of the average wage with the burden faced by their counterparts earning
five-thirds of the average wage (see OECD 2010b:131). Column 4 is compared to
column 5. In 1997, a South African worker earning two-thirds of the average wage
(i.e. the ‘poor’ taxpayer) paid 33.2% of the tax burden faced by the worker earning
five-thirds of the average wage level (the ‘rich’ taxpayer). In 2000, the tax burden of
the lower-paid worker was 41.5% of the ‘rich’ taxpayer’s burden, and in 2009 this
ratio was 39.4%. Overall personal income tax was less progressive in 2009 compared
to 1997, but when the within-time period is considered, personal income taxes first
became less progressive (between 1997 and 2000) and then increased marginally in
progressivity (between 2000 and 2009). The increased progressivity in the last decade
is also observed in OECD countries where the worker earning two-thirds of the
average wage level paid 53% in 2000 and 48% in 2009 of the tax burden of the worker
earning five-thirds of the average wage level (OECD 2010b: 131). In eight OECD
countries, personal income tax became less progressive (including Korea, Mexico
and Poland), and in 20 countries personal income taxes became more progressive.
The distribution of personal income and PIT reforms
According to the Income and Expenditure of Households 2005/06 Survey (IES)
of Statistics South Africa (StatsSA), income inequality is confirmed by high Gini
coefficients. Based on disposable income (from work and social grants) for the whole
country, the Gini coefficient was 0.72. If social grants and taxes are excluded, the
Gini coefficient for the whole country would be 0.80 – that is, the reduction of
inequality through redistributive policies reduces the Gini coefficient by 8 percentage
points (StatsSA 2008: 4). The top 10% of the population and the bottom 90% of the
population each account for approximately 50% of household income reported in
IES, with the poorest 20% accounting for less than 1.5% of income (StatsSA 2008:
33).
Measures to redistribute conventionally include social transfers and a progressive
income tax system. When reported income tax in the IES is considered, the income
skewness in South Africa is matched by a similar skewness in personal income tax
paid. Statistics South Africa (2008: 10–11), however, cautions that the income tax
reported in the IES 2005/06 Survey ‘appears low’ and is poorly captured. The most
affluent 10% of households reported 68% of income tax, while the 50% of households
receiving the lowest incomes accounted for just 1.4% of income tax. The average
effective tax rate – defined as the share of income tax within gross income minus
imputed rent – rises from approximately 1% in decile 1 to 9.3% in decile 10, reflecting
1
47 T.J. Steenekamp
the progressive nature of the South African personal income tax system. StatsSA
(2008: 33) nevertheless concludes that taking taxation into account “reveals no
statistically significant impact on inequality”.
In a fiscal incidence study conducted by Van der Berg (2009), a Gini coefficient for
pre-transfer income of 0.69 is reported. This coefficient dropped to 0.52 for income
plus benefits and to 0.47 after taxes had also been subtracted. Van der Berg (2009:
17) concluded that this illustrated that the South African fiscal process was highly
distributive and that social spending in particular reduced the Gini coefficient much
more than the progressive income tax system did.
Nyamongo and Schoeman (2007) also considered the redistributive effect
of the tax burden in South Africa. Using the Musgave and Thin approach, their
results indicate a small decline in redistribution between 1994 and 1999/2000,
but a significant deterioration between 1999/2000 and 2003/2004. In contrast, the
Kakwani redistributive measure shows a marginal decline in progressivity since 1994
(Nyamongo & Schoeman 2007: 489).
The empirical evidence on the impact of taxation on redistribution in South
Africa since 1994 is not very encouraging. This conclusion is confirmed by other
fiscal incidence studies that show that the redistributive effects of taxes are minor in
developing countries and that expenditure programmes are much more progressive
(see Bird & Zolt 2005: 931). For example, Barreix and Roca (2007: 137–138) show that
in Latin America, the PIT improves the Gini coefficient by more than 1 percentage
point in the case of Mexico only. Unsurprisingly, Bird and Zolt (2008: 81) state
that “the most effective way to reduce inequality is not through taxation but rather
through spending programs targeted at the poor”.
PIT rate changes and thoughts on reforming personal
income tax
Improving the income distribution and more (not less) tax revenue to meet everincreasing demands for public services remain challenges that South Africa faces.
Conventional wisdom, or good tax policy, requires a reduction in PIT rates and of
course that tax collection be further improved and the tax base broadened. PIT rate
changes will be considered here. Another perspective that may shed light on the
revenue and distributional challenges is the possibility of system failure. In other
words, should the tax system not be redesigned to meet these and other challenges?
A comprehensive evaluation of the present system cannot be attempted here, but brief
attention will be given to current thinking on possible alternatives.
1
48 Personal income tax reform in South Africa since 1994
PIT rate changes
Because of the recent global economic crisis, the OECD (2010a: 98) notes that further
reductions in the top PIT rate will be resisted and remarks that many countries are
actually considering raising their top PIT rates. In the UK, the Liberal Democrats
called for a new and draconian tax on bank bonuses (Guardian 2010), and even
billionaire investor Warren Buffet favours the United States taxing the wealthy more
(Timeslive 2010). The OECD (2010a: 98) suggests that a second-best solution to
promote entrepreneurship might be the implementation of an increase in the top
marginal PIT rate accompanied by more generous loss-offset provisions to increase
risk-taking by risk-averse entrepreneurs (OECD 2010a: 98). When developing
countries face revenue challenges, the politically most acceptable approach is also to
raise tax rates. Bird (2008a: 10–11) cautions against this option when the tax base is
small and also politically weak. Horizontal and vertical inequity may be exacerbated,
and the efficiency cost created by tax distortions increases (more or less with the
square of the tax rate).
Another angle on increasing (or decreasing) PIT rates comes from the association
between company income tax (CIT) rates and personal income tax rates. Without
going into the whole debate concerning tax arbitrage opportunities where a gap
between the top PIT and CIT rate occurs, it would suffice to say that low company
tax rates increase the cost of maintaining high personal rates (Ganghof 2006). If
company tax rates are assumed to be exogenous, the results from a simple crosssectional regression analysis of 38 OECD and EU member countries suggest that
lower company tax rates pull down PIT rates on high wages and above-normal
capital income (Ganghof 2006).
In concert with global tax practice, PIT rates in South Africa have reduced from
time to time since 1994. The top marginal personal income tax rate was lowered
in steps from 45% to 42% in 2000/01, and to 40% in 2002/03. Figure 3 compares
the statutory top marginal personal income tax rates in South Africa with those of
countries of the Southern African Development Community (SADC) for 2009. The
average top marginal PIT rate in SADC (30%) is much lower than in the OECD
(41%) and in South Africa (40%). In SADC, the only country where the top marginal
personal income tax rate exceeds that of South Africa is the Democratic Republic of
the Congo (DRC). The PIT rate in South Africa is therefore at the same level as the
average OECD PIT rates. The scope for higher personal income tax rates seems to be
constrained by already-high marginal tax rates, a small tax base and a CIT rate that
is lower than the top marginal PIT rate.
1
49 T.J. Steenekamp
1
OECD Average
SADC Average (excl SA)
Zimbabwe
Zambia
Tanzania
Swaziland
South Africa
Seychelles
Namibia
Mozambique
Mauritius
Malawi
Madagascar
Lesotho
DRC
Botswana
Angola
41
29
35
35
30
33
40
15
37
32
15
30
22
35
50
25
17
0
10
20
30
40
50
60
Top
41PIT rate
Notes:
• Unweighted averages for OECD are the top statutory tax rates (combined central and sub-central) that apply at the threshold level.
• Tax rates for SADC are for central government.
Sources: SADC (2011); OECD (2011); OECD (2010b)
Figure 3: Statutory top marginal PIT rates (2009)
Alternative PIT designs
Given the current thinking and empirical evidence on personal income tax policy
and reform, the direction of fundamental personal income tax reform globally can
be categorised into four alternative personal income tax approaches: reforming the
comprehensive personal income tax (CPIT) system, the flat tax system, presumptive
taxation and the dual income tax (DIT) system (see IFS 2010, 2011; OECD 2006,
2010; Bahl & Bird 2008, Bird 2008a, 2008b; Keen, Kim & Varsano 2008; Bird &
Zolt 2005, 2008, 2010). In brief, the CPIT combines a progressive rate schedule for
all sources of income with a system of tax reliefs; the flat tax system provides for a
single tax rate with a basic tax allowance; presumptive taxation is an administrative
1
50 Personal income tax reform in South Africa since 1994
assessment system using indicators such as assets or turnover; and the DIT combines
a single (low) tax rate on capital income with a progressive rate structure for labour
income.
The income tax system in South Africa conforms to a semi-comprehensive income
tax system. The semi-CPIT system is prone to tax arbitrage as individuals restructure
their tax affairs to exploit exemptions, allowances and (savings and investment) aftertax rate differentials. Globalisation forces least developed countries (LDCs) to lower
tax rates on mobile resources (for example, professionals). Tax revenue comes under
pressure, and both horizontal and vertical equity principles are violated. It is also
a complex system. In the past two decades, a number of alternative approaches to
reforming personal income tax were proposed. One option is to make the CPIT
system better by a further reduction in rates, broadening the labour and capital tax
base, more withholding, even fewer rates (two to three) and allowing expensing of
costs incurred in generating income (including expensing of costs on cash saved and
invested) (see OECD 2010a: chapter 5; IFS 2011). Bird and Zolt (2005: 938) note that
it is far from certain that the expected additional revenue and other benefits expected
from these reforms “will justify the additional administrative and political costs”.
The flat rate income tax option is a response mainly to the administrative
complexity of the comprehensive income tax and the inefficiencies created by taxing
labour income and capital income at different rates (see Barreix & Roca 2007: 127;
Keen et al. 2008).1 Horizontal equity is promoted, but a movement away from a
progressive system will imply large revenue losses and an increase in the tax burden
of low-income and middle-income taxpayers (to maintain revenue neutrality),
jeopardising vertical equity (see Keen et al. 2008: 736–737). The third option for
personal income tax reform is a presumptive tax approach (see Taube & Tadesse
1996; Balestrino & Galmarini 2005; Bird & Zolt 2005). It is an effective way of raising
revenue and broadening the tax base whenever the government lacks the capability
of properly administering an income tax, and horizontal and vertical equity may be
enhanced (by taxing the informal sector and the self-employed). At most, however,
it serves a complementary role to the CPIT system in taxing the hard-to-tax and
drawing this group of informal income earners into the tax net.
The dual income tax (DIT) approach is the “more promising alternative”
suggested for reforming personal income tax in developing countries (see Bird &
Zolt 2005: 940; Bird & Zolt 2010; Boadway 2005).2 The dual income tax system taxes
personal capital income (for example, interest, capital gains, profits from personal
businesses) separately from labour income. A single proportional rate for capital
income is set equal to the lowest income tax rate (the first income tax bracket) on
personal (wage) income. Labour income (for example, wages, salaries) is subject to
51 T.J. Steenekamp
a progressive rate structure, which provides for tax deductions and exemptions to
achieve equity objectives. Overall, the taxation of labour income is therefore higher
than on capital income. A lower flat tax rate on capital income and a progressive tax
on labour income is a pragmatic way to deal with problems of capital flight and tax
arbitrage activities, and reduces the distortions implicit to the differential treatment of
different sources of capital income. By lowering the tax rate on capital, the capital tax
base can be expanded by eliminating tax incentives. The tax system is simplified, and
compliance may improve with concomitant revenue effects. Horizontal inequities are
avoided, and discrimination against savings (using the life-time income approach) is
reduced. Because the tax base is broadened (in that there are no tax preferences and
exclusions), the DIT system is administratively simplified.
Income-splitting mechanisms with all kinds of special tax rules are, however,
required for self-employed and other small entrepreneurs, which makes the DIT
imperfect. This administrative challenge is the Achilles heel of the system (Sørensen
2007: 566; Sørensen 2005: 780–781). It is a matter of concern that adopting different
tax rates for labour and capital income will have revenue implications. If lower rates
are levied on capital income, tax rates on labour income will have to be increased to
maintain or increase income tax revenue. This will have equity consequences. From
a vertical equity perspective, lower tax rates on capital may be viewed as unfair, since
capital income is earned mainly by the rich. Introducing a dual income tax system
in developing countries will not be problem free. A number of design issues emerge
(see Bird & Zolt 2010: 197–206; Barreix & Roca 2007). Policy-makers have to consider
the relative tax rates for labour (i.e. the top marginal PIT rate) and capital income
(for example, equal to or higher than the lowest PIT rate), the zero bracket amount
for labour and capital, the definition of capital income (active and passive capital
income), setting the company tax rate (for example, equal to or higher than the rate
for unincorporated business profits) and taxing normal returns on capital (or only
excess returns).
The DIT provides the flexibility that developing countries need both to meet
the international competition for capital and to maintain or even increase the
progressiveness generated by the personal income tax system. A progressive dual
income tax (and CPIT) satisfies the vertical equity tax principle ‘more accurately’
than a flat tax rate system, and once the life-time income approach is used as the basis
of comparison, the DIT is also horizontally fair (OECD 2006: 131). The Uruguayan
version of a dual income tax shows a more than 2 percentage point improvement in the
after-tax Gini coefficient (Barreix & Roca 2007: 133). This exceeds the disappointing
results obtained using the comprehensive income tax system. The dual income tax
system appears to sacrifice neutrality; income from capital is taxed differently to
52 Personal income tax reform in South Africa since 1994
income from labour. Furthermore, vertical equity is sacrificed for horizontal equity.
The real advantage of the DIT is that efficiency gains can be achieved by taxing all
capital at the same low rate.
Conclusions
Since 1994, the personal income tax system in South Africa has been reviewed and
reformed in many respects. All these reforms were aligned with government’s tax
policy, which aimed to improve collection efficiency, reduce the economic distortions
associated with the tax structure and lower the costs of investment and job creation,
and release household spending power. The approach taken by National Treasury
was in line with the ‘holy trinity’ of tax reform suggested by tax experts (including
the Katz Commission), and required improving tax administration, broadening the
tax base and lowering tax rates.
From the preceding sections, it is evident that before 1994, the share of personal
income tax in total tax revenue relative to other tax sources was at a very high level.
Since 1994, the relative contribution decreased somewhat but remains high. When
the incidence of VAT, the fuel levy and user charges are added and the limited size
of the tax base is considered, it is doubtful whether there is room to increase the
burden. The total tax burden on individuals has remained more or less constant
since 1994 and is high in comparison with other developing countries. Moreover, the
South African PIT base is already taxed very intensively compared with developing
countries. The personal income tax structure is progressive. Between 1997 and 2000,
personal income taxes first became less progressive and then increased marginally
in progressivity between 2000 and 2009. Unfortunately personal income tax reforms
over the last two decades have not really resulted in an improvement of the income
distribution. Furthermore, there is some consensus that personal income taxation is
not a very suitable instrument for redistribution purposes in developing countries.
South Africa, however, still faces the twin challenges of improving the income
distribution and increasing tax revenue to meet ever-increasing demands for public
services. Conventional wisdom, or good tax policy, dictates a reduction in PIT rates,
but as part of the fall-out of the global financial crisis (and for equity reasons), further
reductions will be resisted. Instead, globally there is populist sentiment in favour of
raising top marginal PIT rates. The PIT rate in South Africa is at the same level as
the average OECD rates. The scope for higher personal income tax rates seems to be
constrained by already-high marginal tax rates, a small tax base and a CIT rate that
is lower than the top marginal PIT rate.
1
53 T.J. Steenekamp
Consideration was given to the possibility that the revenue and distributional
challenges are due to the failure of the semi-comprehensive income tax system to
which South Africa adheres. An in-depth analysis of this possibility is simply beyond
the scope of this study, but current international thinking on tax reform by eminent
policy researchers and recent tax reform studies by the Institute for Fiscal Studies
(IFS) and the OECD provided a foundation for a brief discussion of alternative PIT
design options. The tax reform option that holds promise for South Africa (and
developing countries) is the dual income tax system. The DIT provides the flexibility
that developing countries need both to meet the international competition for capital
and to maintain or even increase the progressiveness generated by the personal
income tax system. In designing tax systems, it is almost inevitable that governments
will be faced with policy conflicts and goal choices. Tax fairness, efficiency, neutrality
and administrative simplicity cannot always be achieved simultaneously and must
sometimes be traded off. The dual income tax system appears to sacrifice neutrality,
since income from capital is taxed differently to income from labour. It does have the
advantage of improving horizontal equity.
The forces that impact on tax reform are clearly diverse, but Norregaard and
Kahn (2007: 3–7) suggest that a common set of factors steers countries to adopt the
same approach to tax reform. These include globalisation (and the increased mobility
of factors of production), employment creation and the need to remove labour tax
impediments, a ‘subtle’ change of views in favour of tax efficiency in the efficiency–
equity trade-off and ‘herd behaviour’, involving the adoption of fashionable tax
policies (for example, flat taxes). The dual income tax may well become a popular
alternative to the present comprehensive income tax system in developing countries.
A much more comprehensive analysis is required, however, to determine whether
South Africa should run with the herd.
Acknowledgements
The author would like to thank Prof. Estian Calitz and two anonymous referees for
valuable comments and suggestions. Any remaining errors of fact or interpretation
are the responsibility of the author.
1
Endnotes
1 The sections on the flat rate income tax and DIT draw to some extent from my contribution to income tax reform in Black, Calitz & Steenekamp (2011: 234–239).
54 Personal income tax reform in South Africa since 1994
2 For a comprehensive analysis of the dual income tax system in developed countries, see
Sørensen (2005: 777–801), Sørensen (2007: 557–602) and Sørensen and Johnson (2010:
179–235). The potential for introducing the dual income tax in developing countries is
discussed in great detail in Bird & Zolt (2010) and Boadway (2005: 910–927).
References
Bahl, R.W. & Bird, R.M. 2008. ‘Tax policy in developing countries: Looking back – and
forward’, National Tax Journal, LXI(2): 279–301.
Balestrino, A. & Galmarini, U. 2005. On the redistributive properties of presumptive
taxation. Cesifo Working Paper No. 1381, Category 1: Public Finance.
Barreix, A & Roca, J. 2007. ‘Strengthening a fiscal pillar: The Uruguayan dual income tax’,
CEPAL Review, 92(August): 121–140.
Bird, R.M. 2008a. Tax challenges facing developing countries. Working Paper Series IIB,
Paper No. 12. Institute for International Business. Rotman School of Management,
University of Toronto.
Bird, R.M. 2008b. The BBLR approach to tax reform in emerging countries. International
Studies Program Working Paper 08-04. Andrew Young School of Policy Studies,
Atlanta.
Bird, R.M. & Zolt, E.M. 2005. ‘The limited role of the personal income tax in developing
countries’, Journal of Asian Economics, 16: 928–946.
Bird, R.M. & Zolt, E.M. 2008. ‘Tax policy in emerging countries. Environment and Planning
C’, Government and Policy, 26: 73–86.
Bird, R.M. & Zolt, E.M. 2010. ‘Dual income taxation and developing countries’, Columbia
Journal of Tax Law, 1: 174–217.
Black, P., Calitz, E. & Steenekamp, T.J. 2011. Public economics, 5th edition. Cape Town:
Oxford University Press.
Boadway, R. 2005. ‘Income tax reform for a globalized world: The case for a dual income
tax’, Journal of Asian Economics, 16: 910–927.
Ganghof, S. 2006. The politics of tax structure. Max Planck Institute for the Study of
Societies. MPIfG Working Paper 06/1, January 2006 [Online]. Available at http://www.
mpifg.de/pu/workpap/wp06-1/wp06-1.html Accessed: 4 June 2011.
Guardian. 2010. Liberal Democrats call for draconian tax on bank bonuses. [Online]
Available at http://www.guardian.co.uk/business/2010/oct/05/lib-dem-oakeshott-taxbank-bonuses. Accessed: 6 October 2010.
IFS (Institute for Fiscal Studies). 2010. Dimensions of Tax Design: The Mirrlees Review.
Oxford: Oxford University Press.
IFS 2011. Tax by Design. Oxford University Press. [Online] Available at http://www.ifs.org.
uk/mirrleesReview/design. Accessed: 12 May 2011.
IMF (International Monetary Fund). 2008. Government Finance Statistics Yearbook, 2008,
XXXII. Washington DC: IMF.
55 T.J. Steenekamp
Kakwani, N.C. 1977. ‘Measurement of tax progressivity: An international comparison’,
Economic Journal, 87: 71–80.
Katz Commission. 1994. Report of the commission of enquiry into certain aspects of the tax
structure of South Africa.Pretoria: Government Printer.
Keen, M., Kim, Y. & Varsano, R. 2008. ‘The “flat tax(es)”: Principles and experience’,
International Tax and Public Finance, 15: 712–751.
OECD (Organisation for Economic Cooperation and Development). 2006. Fundamental
Reform of Personal Income Tax. Tax Policy Studies, No. 13. Paris: OECD.
OECD. 2010a. Tax Policy Reform and Economic Growth. Tax Policy Studies, No. 20. Paris:
OECD.
OECD. 2010b. Taxing Wages 2008–2009. Paris: OECD.
OECD. 2011. Tax database. OECD. [Online] Available at: www.oecd.org/ctp/taxdatabase.
Accessed: 5 May 2011.
Manuel, T.A. 2002. The South African tax reform experience since 1994. Address by the
Honourable Trevor A. Manuel, MP, Minister of Finance, Annual Conference of the
International Bar Association. 24 October. [Online] Available at: http://www.treasury.
gov.za/comm_media/speeches/2002/2002102501.pdf. Accessed: 15 May 2011.
Musgrave, R.A. & Thin, T. 1948. ‘Income tax progression’ 1929–48’, Journal of Political
Economy, 56: 498–514.
National Treasury and Department of Finance (various issues). Budget Review. Pretoria:
National Treasury.
National Treasury. 2011. Budget Review 2011. Pretoria: National Treasury.
National Treasury and SARS (South African Revenue Service). 2011. Tax Statistics 2010.
Pretoria: National Treasury.
Norregaard, J. & Khan, T.S. 2007. Tax policy: Recent trends and coming challenges. IMF
Working Paper, WP/07/274. Washington DC: IMF.
Nyamongo, M.E. & Schoeman, N.J. 2007. ‘Tax reform and the progressivity of personal
income tax in South Africa’, South African Journal of Economics, 75(3): 478–495.
SADC (Southern African Development Community). 2011. Tax database. [Online].
Available at: http://www.sadc.int/tifi/tax/chapter/1/country/. Accessed: 14 April 2011.
Sørensen, P.B. 2005. ‘Neutral taxation of shareholder income’, International Tax and Public
Finance, 12: 777–801.
Sørensen, P.B. 2007. ‘The Nordic dual income tax: Principles, practices, and relevance for
Canada’, Canadian Tax Journal, 55(3): 557–602.
Sørensen, P.B. & Johnson, S.M. 2010. ‘Taxing capital income: Options for reform in
Australia’, In Proceedings of a Conference of the Melbourne Institute of Applied Economic
and Social Research on Australia’s Future Tax and Transfer Policy.
SARB (South African Reserve Bank). 2011a. Quarterly Bulletin, March. Pretoria: SARB.
SARB. 2011b. Historical macroeconomic time series information. [Online] Available at:
http://www.resbank.co.za. Accessed: 15 December 2011.
56 Personal income tax reform in South Africa since 1994
StatsSA (Statistics South Africa). 1998. Labour statistics: Employment and salaries and
wages (summary). Statistical release P0200. May. Pretoria: Stats SA.
StatsSA 2003. Comparative labour statistics: Survey of employment and earnings in selected
industries. Discussion paper. Pretoria: Stats SA.
StatsSA. 2008. Income and expenditure of households 2005/06: Analysis of results. Report
No. 01-00-01. Pretoria: StatsSA.
StatsSA. 2011. Quarterly employment statistics, December 2010. Statistical release P0277,
March 2011. Pretoria: Pretoria: StatsSA.
Steenekamp, T.J. 2007. ‘Tax performance in South Africa: A comparative study’, Southern
African Business Review, 11(3): 1–16.
Tanzi, V. 1996. Globalization, tax competition and the future of tax systems. IMF Working
Paper WP/96/141. Washington DC: IMF.
Tanzi, V. 2004. ‘Globalization and the need for fiscal reform in developing countries’,
Journal of Policy Modeling, 26: 525–542.
Taube, G. & Tadesse, H. 1996. Presumptive taxation in sub-Saharan Africa: Experiences
and prospects. IMF Working Paper WP/96/5.
Timeslive. 2010. US should tax rich more: Buffet. [Online] Available at http://www.timeslive.
co.za/world/article691627.ece/. Accessed: 6 October 2010.
UN Millennium Project. 2005. Investing in Development: A Practical Plan to Achieve the
Millennium Development Goals. New York.
Van der Berg, S. 2009. Fiscal incidence of social spending in South Africa, 2006. A report
to National Treasury, 28 February 2009. [Online] Available at: http://www.treasury.gov.
za. Accessed: 15 December 2011.
57